Re: Bug#594623: xserver-xorg-video-intel: after upgrade to 2.12.0+legacy1-1 X freeze on gdm start
On 08/28/2010 02:10 PM, Julien Cristau wrote: Could you guys test the driver pointed at http://lists.freedesktop.org/pipermail/intel-gfx/2010-August/007910.html? (probably with the 2.6.32-20 kernel) Situation summary: - the latest intel drivers and the latest kernel kms blacklisting cause hard system freeze to many (most?) i855 owners; - these changes are, unfortunately, worse than what they want to solve; - upstream is working on a solution (slower but more stable) for these kind of problem that affects older chipset and is designed to work with KMS; - a working solution isn't there now. With these premises, shouldn't these changes be reverted, in particular the kernel blacklist? Up to now i've seen complains only from i855 owners so it can be considered to remove only this chipset, leaving blacklisted the older ones. If that helps i can ask a friend of mine to make some test with his i845. P.S.: for the record, at work i've tested another i855 notebook (Acer TM292LCi) that exibits the same hangs reported in this bug. For few week it will be available if we want test new drivers or patched one. Ciao. Cesare. -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4c8022d2@gmail.com
Re: Bug#594623: xserver-xorg-video-intel: after upgrade to 2.12.0+legacy1-1 X freeze on gdm start
On Mon, Aug 30, 2010 at 00:52:40 +0300, Kamen Naydenov wrote:
If I run glxgears X crashes badly - black screen or screen shot and
only SysRq commands works (can't test network access).
OK, I can reproduce a crash when running glxgears.
Program received signal SIGSEGV, Segmentation fault.
0xb760d91e in ?? () from /lib/i686/cmov/libc.so.6
(gdb) bt full
#0 0xb760d91e in ?? () from /lib/i686/cmov/libc.so.6
No symbol table info available.
#1 0xb76107fc in ?? () from /lib/i686/cmov/libc.so.6
No symbol table info available.
#2 0xb7610d2d in realloc () from /lib/i686/cmov/libc.so.6
No symbol table info available.
#3 0x080a9cf3 in Xrealloc (ptr=0xb76e43a0, amount=3077456856)
at ../../os/utils.c:1122
No locals.
#4 0x080a107e in miRectAlloc (pRgn=0x94a2494, n=20) at ../../mi/miregion.c:392
data = value optimized out
#5 0x080a2890 in miAppendNonO (badreg=0xbffbe398, pOverlap=0xbffbe3bc)
at ../../mi/miregion.c:530
pNextRect = value optimized out
#6 miRegionOp (badreg=0xbffbe398, pOverlap=0xbffbe3bc)
at ../../mi/miregion.c:793
bot = value optimized out
numRects = value optimized out
ytop = value optimized out
newSize = value optimized out
prevBand = 10
top = value optimized out
ybot = value optimized out
curBand = 0
r2y1 = value optimized out
oldData = value optimized out
r1y1 = value optimized out
#7 miRegionValidate (badreg=0xbffbe398, pOverlap=0xbffbe3bc)
at ../../mi/miregion.c:1377
half = 1
numRects = 10
ri = 0x94a2480
numRI = 3
sizeRI = value optimized out
i = value optimized out
rit = value optimized out
box = 0x94a2494
riBox = value optimized out
ret = 1
#8 0x0815eafc in miValidateTree (pParent=0x9160e30, pChild=0x93c6210,
kind=VTMap) at ../../mi/mivaltree.c:741
totalClip = {extents = {x1 = 0, y1 = 25, x2 = 1024, y2 = 600},
data = 0x0}
childClip = {extents = {x1 = 0, y1 = 0, x2 = 0, y2 = 0},
data = 0x81effb8}
childUnion = {extents = {x1 = 0, y1 = 0, x2 = 0, y2 = 600},
data = 0x956a650}
exposed = {extents = {x1 = 0, y1 = 0, x2 = 0, y2 = 0},
data = 0x81effb8}
pScreen = 0x91223a0
pWin = 0x93c6210
overlap = 1
viewvals = value optimized out
forward = -1074011240
#9 0x080992f5 in MapWindow (pWin=0x93c6210, client=0x9393990)
at ../../dix/window.c:2671
event = {u = {u = {type = 19 '\023', detail = 0 '\000',
sequenceNumber = 983}, keyButtonPointer = {pad00 = 64421907,
time = 242, root = 14681649, event = 0, child = 0, rootX = 0,
rootY = 0, eventX = 0, eventY = 0, state = 0,
sameScreen = 0 '\000', pad1 = 0 '\000'}, enterLeave = {
pad00 = 64421907, time = 242, root = 14681649, event = 0,
child = 0, rootX = 0, rootY = 0, eventX = 0, eventY = 0,
state = 0, mode = 0 '\000', flags = 0 '\000'}, focus = {
pad00 = 64421907, window = 242, mode = 49 '1', pad1 = 6 '\006',
pad2 = 224 '\340', pad3 = 0 '\000'}, expose = {pad00 = 64421907,
window = 242, x = 1585, y = 224, width = 0, height = 0,
count = 0, pad2 = 0}, graphicsExposure = {pad00 = 64421907,
drawable = 242, x = 1585, y = 224, width = 0, height = 0,
minorEvent = 0, count = 0, majorEvent = 0 '\000',
pad1 = 0 '\000', pad2 = 0 '\000', pad3 = 0 '\000'},
noExposure = {pad00 = 64421907, drawable = 242, minorEvent = 1585,
majorEvent = 224 '\340', bpad = 0 '\000'}, visibility = {
pad00 = 64421907, window = 242, state = 49 '1', pad1 = 6 '\006',
pad2 = 224 '\340', pad3 = 0 '\000'}, createNotify = {
pad00 = 64421907, parent = 242, window = 14681649, x = 0, y = 0,
width = 0, height = 0, borderWidth = 0, override = 0 '\000',
bpad = 0 '\000'}, destroyNotify = {pad00 = 64421907,
event = 242, window = 14681649}, unmapNotify = {
pad00 = 64421907, event = 242, window = 14681649,
fromConfigure = 0 '\000', pad1 = 0 '\000', pad2 = 0 '\000',
pad3 = 0 '\000'}, mapNotify = {pad00 = 64421907, event = 242,
window = 14681649, override = 0 '\000', pad1 = 0 '\000',
pad2 = 0 '\000', pad3 = 0 '\000'}, mapRequest = {
pad00 = 64421907, parent = 242, window = 14681649}, reparent = {
pad00 = 64421907, event = 242, window = 14681649, parent = 0,
x = 0, y = 0, override = 0 '\000', pad1 = 0 '\000',
pad2 = 0 '\000', pad3 = 0 '\000'}, configureNotify = {
pad00 = 64421907, event = 242, window = 14681649,
aboveSibling = 0, x = 0, y = 0, width = 0, height = 0,
borderWidth = 0, override
Re: Bug#594623: xserver-xorg-video-intel: after upgrade to 2.12.0+legacy1-1 X freeze on gdm start
On Sat, Aug 28, 2010 at 15:10, Julien Cristau jcris...@debian.org wrote: On Sat, Aug 28, 2010 at 00:36:31 +0200, Cesare Leonardi wrote: I've also tried (don't know if it make sense) with 2.6.34-1~expermental.2, always with i915.modeset=0: same freeze. OK, it's starting to sound like this 'legacy' experiment is a failure so far. On my 945GM, legacy crashes on UMS on gnome startup (gdm is ok) reporting a lockup. I don't have an i8xx board to play with. Could you guys test the driver pointed at http://lists.freedesktop.org/pipermail/intel-gfx/2010-August/007910.html? (probably with the 2.6.32-20 kernel) I'll try it if you guide me how to compile and test it, or give me links where to read how to test. I already downloaded xf86-video-intel-shadow.tar.gz but I cant figure how to check dependencies, compile it and to test it best regards Kamen -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktimnqypqolb-crlqgccnlu7fyyfh5dy4z=pn9...@mail.gmail.com
Re: Bug#594623: xserver-xorg-video-intel: after upgrade to 2.12.0+legacy1-1 X freeze on gdm start
On Sun, Aug 29, 2010 at 13:25:07 +0300, Kamen Naydenov wrote: On Sat, Aug 28, 2010 at 15:10, Julien Cristau jcris...@debian.org wrote: Could you guys test the driver pointed at http://lists.freedesktop.org/pipermail/intel-gfx/2010-August/007910.html? (probably with the 2.6.32-20 kernel) I'll try it if you guide me how to compile and test it, or give me links where to read how to test. I already downloaded xf86-video-intel-shadow.tar.gz but I cant figure how to check dependencies, compile it and to test it On Sun, Aug 29, 2010 at 13:17:39 +0200, dmso...@edu.xunta.es wrote: When I have time, I'll try to compile your shadow version and see what happens. Is there a .deb package of the shadow version? It would be easier to try. There are i386 packages at http://people.debian.org/~jcristau/intel-shadow/ which work for me on minimal testing. Although it looks like they crash the X server in dri2 code when I run glxinfo, so I put xserver packages with a fix for that crash at the same place. If you can see if that's more stable than the stock unstable packages when running on kms that would be good. You'll need a Device section in xorg.conf such as: Section Device Identifier intel Option shadow EndSection sha1sums for those files: 7434f94eeff53742888db5747cdfb4efe3db2a24 xserver-xorg-video-intel-dbg_2.12.0-2_i386.deb 6e232126f89b2759ef095135c35070c5432ebbfd xserver-xorg-video-intel_2.12.0-2_i386.deb ad83a07bd714d1b9429a3c087a83b8297f2a21f1 xdmx-tools_1.7.7-5_i386.deb ccc4117d74c88b5d3ae6ccb7b426b65de92f4611 xdmx_1.7.7-5_i386.deb 649fcf10c7bf9943802e018b4d0f2d078ec23af3 xnest_1.7.7-5_i386.deb 942870f5eba5eef813ab13a95efc1b2be93db2e8 xorg-server_1.7.7-5_i386.changes b6d97bb80f41f18b4ef1a9548f1271fea00ad126 xserver-common_1.7.7-5_all.deb f339e415a741e0dba64dff3a50f4973d4292f2fe xserver-xephyr_1.7.7-5_i386.deb dea987480865749bc5292f31d3f391aa91f05cb7 xserver-xfbdev_1.7.7-5_i386.deb 6fa64fb8d406e44a30041080a7c2e71abb496efa xserver-xorg-core-dbg_1.7.7-5_i386.deb f9c21349b0a075f21cb4c74d5d2fe0f6f14fffb1 xserver-xorg-core-udeb_1.7.7-5_i386.udeb 4c86e5cdf9481844d2f48d38eb4894bb03c145ef xserver-xorg-core_1.7.7-5_i386.deb 83d12f4daaf4ff78332f1898e2382febac9d57b7 xserver-xorg-dev_1.7.7-5_i386.deb 2baed634647ec3b7c6499ca945e7ed6219d44fa0 xvfb_1.7.7-5_i386.deb Cheers, Julien signature.asc Description: Digital signature
Re: Bug#594623: xserver-xorg-video-intel: after upgrade to 2.12.0+legacy1-1 X freeze on gdm start
On 08/29/2010 05:18 PM, Julien Cristau wrote: There are i386 packages at http://people.debian.org/~jcristau/intel-shadow/ which work for me on minimal testing. Hi Julien. As long as you provide .deb packages, i can test whatever you want! ;-) Here is my results. Now i'm writing with: - i855 chipset - kernel 2.6.32-20 - KMS enabled with i915.modeset=1 (/etc/modprobe.d/i915-kms.conf temporarly deleted) - intel 2.12.0-2 with Chris Wilson's shadow patch, packaged from you and downloaded from http://people.debian.org/~jcristau/intel-shadow/ - a hand-created file /etc/X11/xorg.conf.d/intel-shadow.conf, containing: Section Device Identifier intel Option shadow EndSection If that helps, i have some packages from experimetal, notably these: firmware-linux-free 2.6.35-1~experimental.2 linux-base 2.6.35-1~experimental.2 linux-image-2.6.34-1-6862.6.34-1~experimental.2 linux-image-2.6.35-rc6-686 2.6.35~rc6-1~experimental.1 linux-image-2.6.35-trunk-6862.6.35-1~experimental.2 libdrm-intel1 2.4.21-1 libdrm2 2.4.21-1 libgl1-mesa-dri 7.8.2-2 libgl1-mesa-glx 7.8.2-2 libglu1-mesa-glx7.8.2-2 Since 2.6.32-21 has KMS disabled in the code, with the configuration above the X server doesn't start: no hangs, only an error reporting that the X server cannot be started and a message in the log reporting that KMS is required (see the X shadow (KMS off).log). So i've downgraded the kernel to 2.6.32-20 (from testing) and downloaded your debs to do some tests. Even if you reported hangs updating only the intel driver, i decided to start step-by-step, also because we have different chipsets. So i've installed only xserver-xorg-video-intel_2.12.0-2_i386.deb and rebooted with the i915.modeset=1 kernel parameter. Here everything seems to works fine so far (uptime 2,30) and i haven't noted any visible slowdown in the day-to-day use under Gnome: web browsing with flash, mail, music, movies, ecc. I'm only not completely sure the patch is working properly: the X log (see X shadow (KMS on).log) doesn't show any error about the shadow option, but doesn't neither show the Shadow buffer allocated message that Chris Wilson said we should expect: http://lists.freedesktop.org/pipermail/intel-gfx/2010-August/007924.html So However i've rapidly rebooted (only up to the GDM prompt) to see that the driver works even with 2.6.34, 2.6.35-rc6 and 2.6.35 (the latter yet without visible mouse pointer, as reported in #592415). I'm available to test every new update of the intel driver and i've taken the liberty to cc Chris Wilson to make him aware of these tests. Ciao. Cesare. --- X log with shadow patch and 2.6.32-21 (KMS blacklisted in kernel code) [...] (EE) open /dev/fb0: No such file or directory (EE) intel(0): No kernel modesetting driver detected. (II) UnloadModule: intel (EE) Screen(s) found, but none have a usable configuration. Fatal server error: no screens found --- X log with shadow patch and 2.6.32-20 (KMS enabled with i915.modeset=1) X.Org X Server 1.7.7 Release Date: 2010-05-04 X Protocol Version 11, Revision 0 Build Operating System: Linux 2.6.26-2-amd64 i686 Debian Current Operating System: Linux tommaso 2.6.32-5-686 #1 SMP Thu Aug 12 13:38:27 UTC 2010 i686 Kernel command line: BOOT_IMAGE=/boot/vmlinuz-2.6.32-5-686 root=UUID=07b633c9-1242-48ee-a6f4-f37c43f639fd ro quiet lapic hpet=force i915.modeset=1 Build Date: 24 August 2010 02:59:40PM xorg-server 2:1.7.7-4 (Julien Cristau jcris...@debian.org) Current version of pixman: 0.16.4 Before reporting problems, check http://wiki.x.org to make sure that you have the latest version. Markers: (--) probed, (**) from config file, (==) default setting, (++) from command line, (!!) notice, (II) informational, (WW) warning, (EE) error, (NI) not implemented, (??) unknown. (==) Log file: /var/log/Xorg.0.log, Time: Sun Aug 29 21:02:50 2010 (==) Using config directory: /etc/X11/xorg.conf.d (==) Using system config directory /usr/share/X11/xorg.conf.d (==) No Layout section. Using the first Screen section. (==) No screen section available. Using defaults. (**) |--Screen Default Screen Section (0) (**) | |--Monitor default monitor (==) No device specified for screen Default Screen Section. Using the first device section listed. (**) | |--Device intel (==) No monitor specified for screen Default Screen Section. Using a default monitor configuration. (==) Automatically adding devices (==) Automatically enabling devices (WW) The directory /usr/share/fonts/X11/cyrillic does not exist. Entry deleted from font path. (WW) The directory /usr/share/fonts/X11/100dpi/ does not exist. Entry deleted from font path. (WW) The directory /usr/share/fonts/X11/75dpi/ does not exist. Entry deleted from font path. (WW) The directory /usr/share/fonts/X11/100dpi does not exist. Entry deleted from font path. (WW) The
Re: Bug#594623: xserver-xorg-video-intel: after upgrade to 2.12.0+legacy1-1 X freeze on gdm start
On Sun, Aug 29, 2010 at 18:18, Julien Cristau jcris...@debian.org wrote: There are i386 packages at http://people.debian.org/~jcristau/intel-shadow/ which work for me on minimal testing. Although it looks like they crash the X server in dri2 code when I run glxinfo, so I put xserver packages with a fix for that crash at the same place. If you can see if that's more stable than the stock unstable packages when running on kms that would be good. You'll need a Device section in xorg.conf such as: Section Device Identifier intel Option shadow EndSection Thank you Julien, for packaging of shadow branch and relevant X packages. I've installed only: xserver-xorg-video-intel_2.12.0-2_i386.deb (with this installed my X crashes on glxinfo) xserver-xorg-core_1.7.7-5_i386.deb xserver-common_1.7.7-5_all.deb Created /etc/X11/xorg.conf with section described above. I have options i915 modeset=1 in /etc/modprobe.d/i915-kms.conf and linux kernel 2.6.32-5-686 (Debian 2.6.32-20) For tests I'm runing X session started via startx. With this setup I can crash (drop me to tty from which I've run startx) X when I run pinball or foobillard. I'm attached glxinfo output and Xorg.0.log (crashed with starting foobillard) When X session is started via GDM crashes just restart gdmgreeter. If I run glxgears X crashes badly - black screen or screen shot and only SysRq commands works (can't test network access). If you need more info I'll provide it. best regards Kamen shadow-logs.tar.gz Description: GNU Zip compressed data
Re: Bug#594623: xserver-xorg-video-intel: after upgrade to 2.12.0+legacy1-1 X freeze on gdm start
For Chris Wilson: you can follow this bug history here: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=594623 On 08/29/2010 05:18 PM, Julien Cristau wrote: Although it looks like they crash the X server in dri2 code when I run glxinfo, so I put xserver packages with a fix for that crash at the same place. Yes, running the glxinfo after updating only the intel-x-driver make the xserver crash, then gdm automatically restarts. In /var/log/X.log.0.old i can see this error: Backtrace: 0: /usr/bin/X (xorg_backtrace+0x3b) [0x80d920b] 1: /usr/bin/X (0x8048000+0x581d5) [0x80a01d5] 2: (vdso) (__kernel_rt_sigreturn+0x0) [0xb786d40c] 3: /usr/lib/xorg/modules/extensions/libdri2.so (0xb785b000+0x1423) [0xb785c423] 4: /usr/lib/xorg/modules/extensions/libdri2.so (0xb785b000+0x1d6a) [0xb785cd6a] 5: /usr/bin/X (0x8048000+0x38067) [0x8080067] 6: /usr/bin/X (0x8048000+0x1e92a) [0x806692a] 7: /lib/i686/cmov/libc.so.6 (__libc_start_main+0xe6) [0xb75a0c76] 8: /usr/bin/X (0x8048000+0x1e511) [0x8066511] Segmentation fault at address (nil) Fatal server error: Caught signal 11 (Segmentation fault). Server aborting I've experienced this with both 2.6.32-20 and 2.6.35-rc6. Julien, many thanks for working on this and providing this experimental packages. I hope we can find a reasonable solution for Squeeze. Ciao. Cesare. -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4c7ae074.1050...@gmail.com
Re: Bug#594623: xserver-xorg-video-intel: after upgrade to 2.12.0+legacy1-1 X freeze on gdm start
On Sat, Aug 28, 2010 at 00:36:31 +0200, Cesare Leonardi wrote: I've also tried (don't know if it make sense) with 2.6.34-1~expermental.2, always with i915.modeset=0: same freeze. OK, it's starting to sound like this 'legacy' experiment is a failure so far. On my 945GM, legacy crashes on UMS on gnome startup (gdm is ok) reporting a lockup. I don't have an i8xx board to play with. Could you guys test the driver pointed at http://lists.freedesktop.org/pipermail/intel-gfx/2010-August/007910.html? (probably with the 2.6.32-20 kernel) Cheers, Julien signature.asc Description: Digital signature
Re: Bug#594623: xserver-xorg-video-intel: after upgrade to 2.12.0+legacy1-1 X freeze on gdm start
Julien Cristau wrote: On Sat, Aug 28, 2010 at 00:36:31 +0200, Cesare Leonardi wrote: I've also tried (don't know if it make sense) with 2.6.34-1~expermental.2, always with i915.modeset=0: same freeze. OK, it's starting to sound like this 'legacy' experiment is a failure so far. On my 965GM it's a huge improvement over the driver in testing which lasts seconds and is unusable. I've been running with it for some time now and only two crashes so far, no accelerated video is the thing I miss most. I'm running on 2.6.34-1 (2.6.35-rc5 was unstable) Regards, M -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/e039223b3e530b16bea7eabd1ebf6e6d.squir...@luminoussheep.net
